Berkeley Artificial Intelligence researchers are joining forces with climate experts, government agencies, and industry, as part of the new Berkeley AI Research (BAIR) Climate Initiative, a multi-disciplinary student-led hub dedicated to fighting climate change.  The effort is being led by co-founding director CS Prof. Trevor Darrel and organized by three of his graduate students, Colorado Reed (co-advised by Kurt Keutzer), who will help lead the initiative, Medhini Narasimhan, and Ritwik Gupta (co-advised by Shankar Sastry).  Their objective is to develop AI techniques that address problems with data processing, particularly involving massive data sets. To maximize the benefit to other researchers studying the same problems around the world, all work done by the initiative will be openly published and available without exclusive or proprietary licensing.  One of their first projects, “The Fate of Snow,” will be a collaboration between BAIR Climate Initiative researchers and other scientists and policy experts on the Berkeley campus, Berkeley Lab (LBNL), Meta AI (which belongs to Meta Platforms, Inc.) and the Center for Western Weather and Water Extremes. The researchers plan to apply AI methods to a multitude of openly available weather and satellite data sources to estimate how much water is in the Sierra Nevada snowpack and forecast what that will mean for streadmflow in the region.

Two of EECS Prof. Ren Ng's former graduate students, Pratul Srinivasan and Benjamin Mildenhall, jointly received an honorable mention for the 2021 Association for Computing Machinery (ACM) Doctoral Dissertation Award.  This award is presented annually to the "author(s) of the best doctoral dissertation(s) in computer science and engineering."  Srinivasan and Mildenhall, who both currently work at Google Research,  were recognized "for their co-invention of the Neural Radiance Field (NeRF) representation, associated algorithms and theory, and their successful application to the view synthesis problem."  Srinivasan’s dissertation, "Scene Representations for View Synthesis with Deep Learning," and Mildenhall’s dissertation, “Neural Scene Representations for View Synthesis,” addressed a long-standing open problem in computer vision and computer graphics called the "view synthesis" problem:  If you provide a computer with just a few of photographs of a scene, how can you get it to predict new images from any intermediate viewpoint?  "NeRF has already inspired a remarkable volume of follow-on research, and the associated publications have received some of the fastest rates of citation in computer graphics literature—hundreds in the first year of post-publication."

EECS Assistant Prof. Nika Haghtalab and CS Assistant Prof. and Associate Prof. in the School of Information, David Bamman, are quoted in a Computing, Data Science, and Society (CDSS) article about the resiliency and determination of the 2022 graduating class, particularly during the pandemic. “This generation of students has persevered, despite these global challenges, to forge a real community with their peers,” said Bamman. They also anticipated the ways the graduates will use their new skills to shape our collective future. “We need graduates who understand the technical methods of data science, their limitations and sources of bias, and the broader context in which information is used to drive policy, inform decision-making, and shape opinion,” Bamman said.  Haghtalab noted that “this is a great time to enter the workforce and contribute to the shaping of data science and computing for the advancement and betterment of the world.”

CS undergraduate student Chase Norman is among 200 young mathematics and computer science researchers selected from across the globe to attend the 9th Heidelberg Laureate Forum (HLF) in Germany this September. During the week-long conference, participants will share ideas with some of the "most exceptional mathematicians and computer scientists of their generations," namely the recipients of some of the field’s most prestigious awards: the Abel Prize, ACM A.M. Turing Award, ACM Prize in Computing (won this year by Berkeley CS Prof. Pieter Abbeel), Fields Medal, and Nevanlinna Prize. Participants and laureates will interact through a blend of scientific and social activities that are designed to foster a relaxed atmosphere and encourage scientific exchange.  Participants are selected by a panel of international reviewers on the basis of their research experience, social engagement skills, and letter of motivation.”. Norman is a CS and Math double major who was admitted to the EECS Honors Program in the breadth area of Mathematical Logic and Foundations.  He is also the president of the CS honor society Upsilon Pi Epsilon, was course staff for CS 170 and CS 61A, and was a percussionist with UC Jazz and the UCB Symphony Orchestra.

CS graduate student Dave Epstein (advisor: Alexei Efros) has won a 2022 Paul & Daisy Soros Graduate Fellowship for New Americans.  This fellowship recognizes outstanding graduate students who are immigrants and children of immigrants in the United States, and "who are poised to make significant contributions to US society, culture or their academic field."  Epstein is affiliated with the Berkeley AI Research (BAIR) Lab where he is teaching machines to solve visual problems without labels, and enabling a creative understanding of the real world to emerge. He is also interested in language, machine learning, synthesis, and interaction. Paul & Daisy Fellowships come with a $90K award.

Three EECS graduate students, Aviral Kumar (advisor: Sergey Levine), Serena Wang (advisors: Rediet Abebe and Michael Jordan), and Eric Wallace (advisors: Dan Klein and Dawn Song) have been named 2022 recipients of the Apple Scholars in AI/ML PhD fellowship.  This fellowship recognizes graduate and postgraduate students in the field of Artificial Intelligence and Machine Learning who are "emerging leaders in computer science and engineering" as demonstrated by their "innovative research, record as thought leaders and collaborators, and commitment to advance their respective fields."  Kumar is working in the area of "Fundamentals of Machine Learning" to develop "reinforcement learning algorithms and tools that enable learning policies by effectively leveraging historical interaction data and understanding and addressing challenges in using RL with deep neural nets." Wang is working in the area of "AI for Ethics and Fairness" to "foster positive long-term societal impact of ML by rethinking ML algorithms and practices, employing tools from robust optimization, constrained optimization, and statistical learning theory."  Wallace is working in the area of "Privacy Preserving Machine Learning," to make "NLP models more secure, private, and robust." Apple Scholars receive support for their research, internship opportunities, and a two-year mentorship with an Apple researcher in their field.

A paper with lead author EECS postdoc Efrat Shimron and co-authors EECS graduate student Ke Wang, UT Austin professor Jonathan Tamir (EECS PhD ’18), and EECS Prof. Michael Lustig shows that algorithms trained using "off-label" or misapplied massive, open-source datasets are subject to integrity-compromising biases.  The study, which was published in the Proceedings of the National Academy of Sciences (PNAS), highlight some of the problems that can arise when data published for one task are used to train algorithms for a different one.  For example, medical imaging studies which use preprocessed images may result in skewed findings that cannot be replicated by others working with the raw data.  The researchers coined the term “implicit data crimes” to describe research results that are biased because algorithms are developed using faulty methodology. “It’s an easy mistake to make because data processing pipelines are applied by the data curators before the data is stored online, and these pipelines are not always described. So, it’s not always clear which images are processed, and which are raw,” said Shimron. “That leads to a problematic mix-and-match approach when developing AI algorithms.”

CS graduate student Chandan Singh (advisor: Bin Yu) has made it to the semi-finals of the 2022 Berkeley Grad Slam Competition, a UC showcase for graduate student research presented in three-minute talks for a general audience, likened to short Ted Talks.  In "Unlocking Scientific Secrets by Distilling Neural Networks," Singh hopes to build on recent advances in machine learning to improve the world of healthcare.   His research focuses on how to build trustworthy machine-learning systems by making them more interpretable through partnerships with domain experts (e.g. medical doctors and cell biologists). These collaborations give rise to useful methodology that both build more transparent models as well as improve the trustworthiness of black-box models. He hopes to help bridge the gap between both types of models so that they can be reliably used to improve real-world healthcare.

CS graduate student Lucas Spangher (advisor: Costas Spanos) gathered musicians from all over the Bay Area to perform a benefit concert in support of Ukraine on March 13th.  Opera and gospel singers, violists, pianists and harpists, were among the more than one dozen volunteers to participate in the Benefit Concert for Humanitarian Aid for Ukraine at Herbst Hall in San Francisco.  Spangher, who plays the cello, reached out to other local musicians on social media to ask if anyone would be interested in participating in an informal, online musical performance in honor of Ukraine, and it expanded from there. “It turned into this amazing professional operation,” said Spangher, “which I think just speaks to the energy and communal desire to do something. This is more than just a fundraiser. It’s a political statement and a way to honor Ukraine’s amazing contributions to classical music that can’t be erased by a vicious autocrat.”  Spangher is a committed climate change activist whose research focuses on how to make artificial intelligence become more flexible for a transition to green energy.  All proceeds from the performance have been donated to Nova Ukraine.

Students in the Data Science Discovery Program are filling a gap in engineering resources to help journalists more easily sort through large stores of records for their research.  The Discovery Program, which is part of Berkeley's Division of Computing, Data Science, and Society (CDSS), connects  around 200 undergraduates with hands-on, team-based data science research projects at Berkeley, government agencies, community groups, and entrepreneurial ventures.  Students have worked on projects like the SF Chronicle's air quality map, the Wall Street Journal's effort to analyze its source and topic diversity using natural processing language, and the California Reporting Project's police misconduct database. “I don’t know if we’d be able to do this without them,” said KQED data reporter Lisa Pickoff-White. “None of these newsrooms would be able to automate this work on their own.”